Load Balancing in the Parallel Queueing Web Server System

Abstract:

With the explosive use of internet, contemporary web servers are susceptible to overloads during which their services deteriorate drastically and often lead to denial of services. Many companies are trying to address this problem using multiple web servers with a front-end load balancer. Load balancing has been found to provide an effective and scalable way of managing the ever-increasing web traffic. Load balancing is one of the central problems that have to be solved in parallel queueing web server system. To analyze load balancing this paper presents a queueing system that has two web servers. Firstly the centralized load balancing system is considered. The next, two routing policies are studied, the average response time and the rejection rate are derived. Finally some of our results are further considered.

• Periodical: Applied Mechanics and Materials (Volumes 143-144)
• Pages: 346-349
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.143-144.346
• Online since: December 2011
• Authors: Li Na Zhang, Xue Si Ma
• Keywords: Load Balancing, Parallel Queues, Routing Control, Web Server

References

[1] F.J. Derfler, L. Freed. Crash-proof your web site. PC Magazine. 19(7) (2000), 135-150.

[2] M. Arlitt, T. Jin. Workload characterization of the 1998 world cup web site, HP Labs Technical Report, Hewlett Packard Palo Alto, CA, September (1999).

[3] M. Arlitt, C.L. Williamson, Web server workload characterization, the search for invariants, in: Proceedings of the 1996 ACM SIGMERICS Conference on Measurement and Modeling of Computer Systems, Philadelphia, PA, May 1996, pp.126-138.

DOI: 10.1145/233013.233034

[4] P. Barford, A. Bestavros, A. Bradley, M.E. Crovella. Changes in Web client access patterns: characteristics and caching implications, World Wide Web 2(Special Issue on Characterization and Performance Evaluation) (1999) 15-28.

DOI: 10.1023/a:1019236319752

[5] P. Barford, M. Crovella. Generating representative Web workloads for network and server performance evaluation, in: proceeding of the 1998 ACM SIGMERICS Conference on Measurement and Modeling of Computer Systems, Madison, WI, June 1998, ACM, New York, pp.151-160.

DOI: 10.1145/277851.277897

[6] M. Harchol-Balter, A. Downey. Exploiting process lifetime distributions for dynamic load balancing, ACM Trans. Comput Syst. 15(3) (1997) 253-285.

DOI: 10.1145/263326.263344

[7] L. Flatto, H.P. McKean. Two queues in parallel. Communications of pure and Applied Mathematics 30(1977), 255-263.

DOI: 10.1002/cpa.3160300206

[8] S. Halfin. The shortest queue problem. Journal of Applied Probability 22(1985), 865-878.

DOI: 10.2307/3213954

[9] B.W. Conolly. The autostrada queueing problem. Journal of Applied Probability 21(1984), 394-403.

DOI: 10.2307/3213648

[10] Y.S. Zheng, P. Zipkin. A queueing model to analyze the value of centralized information. Operations Research 38(2) (1990), 296-307.

DOI: 10.1287/opre.38.2.296